We explore spatial–temporal dynamics in fire regime features at different spatial levels, together with seasonal scales, from historical fire data (1974–2013). Temporal shifts in fire features are investigated by means of change point detection and trend analysis. Our results suggest that most fire features show remarkable shifts between the late 1980s and the first half of the 1990s, with high spatial and intra-annual variability. PCA enabled trends to be synthesized into four main components.

Landslides are a threat not only to people but also to important infrastructure, like highways. Nowadays there are several monitoring systems that are able to detect slope displacements in order to give prompt alarms. On the other hand, such instruments produce a huge amount of information, which is often not totally used and which can also represent an issue for data storage and transmission. In this paper we explain how we dealt with the large quantity of data provided by one of these tools.

The paper presents a simple procedure to select ground-motion time histories with target response spectra for deterministic seismic hazard analysis from the Next Generation Attenuation database using a recently proposed search engine. The recommended time histories can reasonably reflect region-specific seismic hazards and can be used as basic results that are serviceable for common engineering practice.

Assessment of design waves is performed using generalized extreme value (GEV) and generalized Pareto distribution (GPD) based on buoy data for 8 years and ERA-Interim reanalysis data for 38 years. The initial distribution method underestimates return levels compared to GPD. Intercomparison of return levels by block maxima and r-largest method for GEV theory shows that return level for 100 years is 7.24 m by r-largest series. A single storm can cause a large difference in the 100-year H_s value.

A local scale GB-InSAR system was implemented for mapping and monitoring slope landslide residual deformations and for early warning purposes in case of landslide reactivations, with the aim of assuring the safety of the valley inhabitants and the personnel involved in the post-event recovery phase. The here presented methodology could represent a useful contribution to a better understanding of landslide phenomena and decision making process during the post-emergency management activities.

The predictability of 25 historical winter storms over Europe is revisited by taking advantage of a homogeneous dataset of retrospective forecasts for the 1995–2015 period. The forecasts well predict the storms up to 2–4 days ahead only but also show clear potential for the early warning of storms up to 10 days ahead. However, the predictability of individual storms exhibits large variability and physical characteristics are identified for outliers with a poor predictability.

In this study, we have focused on automatically detecting avalanches and classifying them into release zones, tracks, and run-out zones based on aerial imagery using an object-based image analysis (OBIA) approach. We compared the results with manually mapped avalanche polygons, and obtained a user’s accuracy of > 0.9 and a Cohen’s kappa of 0.79–0.85. Testing the method for a larger area of 226.3 km², we estimated producer’s and user’s accuracies of 0.61 and 0.78, respectively.

Changes in lake count and area in the region of Djungarskiy Alatau were assessed, showing that both increased by 6 % in 2002–2014 due to glacier melt. A total of 50 lakes were identified as potentially dangerous. GlabTop2 was used to simulate location and size of overdeepenings in the subglacier beds which present sites where lakes can develop in the future. The model predicted 67 % of lakes would form in the area de-glacierized in 2002–2014, correctly proving a useful tool in hazard management.